WikiResearch on Nostr: "Can Open-Source LLMs Compete with Commercial Models? Exploring the Few-Shot ...
Published at
2024-07-29 10:04:50Event JSON
{
"id": "6e8016fb853fd0b64aa57774f72c2fd71042be81e46bfdc3b08d3554fc503a79",
"pubkey": "3eee297347e2cad49e0095a31aca38ffdcd46865ae75ea8ac5963999ce24d8a1",
"created_at": 1722247490,
"kind": 1,
"tags": [
[
"proxy",
"https://mastodon.social/users/wikiresearch/statuses/112869211569823068",
"activitypub"
]
],
"content": "\"Can Open-Source LLMs Compete with Commercial Models? Exploring the Few-Shot Performance of Current GPT Models in Biomedical Tasks\" commercial VS open-source performance gap exists mainly in zero-shot settings, with Wikipedia context helping in some cases\n\nhttps://t.co/9FoWGnkVPI https://t.co/TBaGXCg4mK\n\nvia https://twitter.com/WikiResearch/status/1817863259686392157",
"sig": "7d29c0119c46fe52c7bbbbf34d11e5a5bfb1ddbffcf7e7c729d1956e76ed48ab9277ca11a24e2d08db7f5157f0b111d3512d12ab42fc95c70249db3ec0fb9f33"
}